During the last fifteen years, the lasso pro cedure has been the target of a substan tial amount of theoretical and applied re search. Correspondingly, many results are known about its behavior for a fixed or op timally chosen smoothing parameter (given up to unknown constants). Much less, how ever, is known about the lasso’s behavior when the smoothing parameter is chosen in a data dependent way. To this end, we give the first result about the risk consistency of lasso when the smoothing parameter is cho sen via crossvalidation. We consider the highdimensional setting wherein the number of predictors p = n,> 0 grows with the